1997 | OriginalPaper | Buchkapitel
Instance Generation
verfasst von : Dr. Alf Kimms
Erschienen in: Multi-Level Lot Sizing and Scheduling
Verlag: Physica-Verlag HD
Enthalten in: Professional Book Archive
Aktivieren Sie unsere intelligente Suche um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
As we have learned, the problem that we are concerned about has not been treated elsewhere. Hence, there is neither an established multi—level test—bed nor an instance generator available. Section 4.1 therefore introduces a parameter controlled instance generator (APCIG) which allows to create multi—level lot sizing (and scheduling) instances systematically. Due to preliminary computational experience, we know that certain parameters of the PLSP have an impact on the performance of solution methods. This is not a very surprising result, of course. But, to gain a better understanding in what makes instances hard to solve and what method should be chosen for what instance, a (full) factorial design is used rather than a randomized design. For a guideline for designing test—beds, performing computational studies, and reporting the results1 we refer to [BaGoKeReSt95, Hoo95]. Section 4.2 presents an experimental design for the PLSP-MM. Section 4.3 relates the speed of different computers to allow fair comparisons with other platforms than those used for our tests.